I completely changed how I teach fisheries and wildlife sciences, and I used Bigfoot to do it. I realized we don’t have to choose between teaching analytical skills or scientific literacy. After all, science is based on the process of inquiry. As researchers, we acknowledge this fact everyday: we have to, we live it.

Isn’t it ironic then, that when it comes to teaching science, we rarely ground our classrooms in the tenets of our profession? Instead, we often provide students a litany of facts, lecture them about methods we think they should know, then expect them to regurgitate the material they have been “taught”. But this isn’t how science works, and teaching science this way denies students the experience of doing and learning science.

There’s a rich body of literature describing how educators can implement inquiry-based learning. One such work, Orchestrating Inquiry Learning, is what I used to change how my classroom works. It says that “inquiry… when supported by technology, can foster higher order thinking skills, and offer learners a meaningful and productive approach to the development of their knowledge of the world.” To be honest, the first time I read that I thought, well, doesn’t that sound nice? Higher order thinking skills? Meaningful and productive? We’ll see.

The first chapter of the book lists the qualities of good inquiry lessons/instruction as:

• Creating an authentic experience
• Student driven
• Not letting the instructor be the expert
• The instructor asks questions and doesn’t provide answers
• Use humor to maintain student interest
• Stress not only the “know how” but also the “know why”

While going down this rabbit hole of inquiry based pedagogy, I was also developing a lesson in how to teach spatial statistics to a classroom of diverse students (by diverse I mean, undergrads and grads in majors ranging from biology to sociology). How could I capture these 6 qualities of good inquiry based lessons, yet still make lessons relevant to my diversity of students?

Bigfoot, that’s how.

Think about it: Bigfoot satisfies all of the above hallmarks of good inquiry based learning.

• It is authentic
• I don’t have pre-derived data sets
• I’m certainly not an expert
• I don’t have the answers
• It’s quite entertaining and comical to be researching predictors of Bigfoot sightings.

The results of this inquiry lesson were also nothing short of entertaining. Below you will find one student’s results and how she used regression techniques to derive a model that can be used to predict Bigfoot sightings on a county-by-county basis in our home state of Ohio (which happens to have the third highest number of Bigfoot sightings in the U.S.).

_________________________

Data: The Bigfoot Field Researchers Organization provided me with a list of Bigfoot sightings across Ohio and the US Census Data a dataset of 47 explanatory variables.

Methods: After reducing the data set using a correlation matrix, I used RStudio and ArcMap to run linear regressions, spatial regressions, geographically weighted regressions (GWR), AIC model selection and a hot spot analysis. Each type of regression is able to give different information about the landscape (Table 1) and then AIC selection will select the model that is most parsimonious. AIC selection works a lot like golf: the lowest AIC score wins.

	Attribute	AICc
Linear Regression	Non-Spatial	209.1
Spatial Regression	Systematic spatial patters	464.36
GWR	Accesses localized patterns	-333.69

I ran several models using variables hypothesized to predict Bigfoot sightings. A total of 50 different models were tested with the number of input explanatory variables ranging from 1 to 15. I used AIC selection to determine the most parsimonious model.

I then conducted a hot spot analysis to see if there were clusters of counties with high or low Bigfoot sightings (spatial heterogeneity of response, ŷ).

Results: The most parsimonious model included the following predictors:

• Number of people that identified as Caucasian
• Number of mobile homes
• Number of individuals from the age of 18 to 29

Of the three regression techniques, the GWR was most parsimonious (Table 1). We found the standardized residuals of the GWR were not spatially autocorrelated, and were able to accurately predict Bigfoot sightings in 60 of the 88 Ohio counties (R²>0.6). This accounts for ~70% of the counties in Ohio. The hot spot analysis showed that there was a cold spot in northwest Ohio and a hot spot in northeast Ohio (Figure 1).

________________________

Conclusion: Inquiry based learning with Bigfoot effectively spurred student interest in learning and exploring spatial statistic techniques, liberated them from preconceived misconceptions about what they should find and encouraged them to seek answers and interpret their results for themselves (there was no Big Foot expert to ask, “does this look right”). By the conclusion of the unit, students had a clear understanding of not only the knowhow of regression and spatial regression approaches, but also the know why of these approaches. Finally, it was really fun to teach this way. It was incredibly rewarding to watch the students learn and explore these techniques and stretch the limits of these approaches.

My only regret is that I may have inadvertently encouraged some students to drop their theses for careers in cryptozoology, but overall I can live with that.

–by Andrew Gregory and Emma Spence, guest bloggers

Dr. Andrew J. Gregory is an Assistant Professor in the School of Earth Environment and Society at Bowling Green State University

Emma Spence is a MS student in Dr. Gregory’s lab at Bowling Green